The good news is that our ability to manage our personal metadata has improved. The bad news is that our ability to work with a shared metadata has not.
It has become a lot easier to manage your personal multimedia collection, your address book and even filing of bills and mails. Virtually all our data can be easily accessed on and synchronized across various devices. But the wheels fall off when you cross to a metadata parallel universe, or in other words, when you switch from one person to another. The way I organize my data is not the same as how you do it.
Now don't get me wrong, I am a big fan of diversity and individualism. The problem is that when you need to collaborate, you have to work out of a common basis which includes semantics and collaboration platforms.
Language, which is one of my favorite examples of a standard is a fundamental tool used by animals and humans to communicate. Without a pre-defined and agreed set of rules it is impossible to communicate. You need to have a shared set of meanings represented by symbols (visual audible etc.)
However, it is also a well-known principle that a degree of freedom is necessary in order for a machine to operate optimally. Now of course, if the parts are too tight, the machine cannot move at all, and if the degree of freedom is too large, excessive wear and tare can occur. So in practice, the degree of freedom is in effect a reflection of the level of control the operator, or manager, of a "machine" decides to apply.
In metadata management specifically, it is not uncommon to apply a wide-brush stoke approach. There is either too much or too little metadata management. This simply indicates that not enough effort has gone in to designing and controlling the metadata management. And the cost? Information risk management, either through opportunity loss, information liabilities and / or productivity loss.
Any good data management program must take into account the impact of doing too little, or too much, in controlling their metadata. Ask yourself how much time is spent resolving issues rooted in misunderstandings? How compatible is your metamodel with external standards?
Semantic coherence and metadata cognition are simply modern terms for the classic story of the tower of babel. And I wonder if we will ever learn...
It has become a lot easier to manage your personal multimedia collection, your address book and even filing of bills and mails. Virtually all our data can be easily accessed on and synchronized across various devices. But the wheels fall off when you cross to a metadata parallel universe, or in other words, when you switch from one person to another. The way I organize my data is not the same as how you do it.
Now don't get me wrong, I am a big fan of diversity and individualism. The problem is that when you need to collaborate, you have to work out of a common basis which includes semantics and collaboration platforms.
Language, which is one of my favorite examples of a standard is a fundamental tool used by animals and humans to communicate. Without a pre-defined and agreed set of rules it is impossible to communicate. You need to have a shared set of meanings represented by symbols (visual audible etc.)
However, it is also a well-known principle that a degree of freedom is necessary in order for a machine to operate optimally. Now of course, if the parts are too tight, the machine cannot move at all, and if the degree of freedom is too large, excessive wear and tare can occur. So in practice, the degree of freedom is in effect a reflection of the level of control the operator, or manager, of a "machine" decides to apply.
In metadata management specifically, it is not uncommon to apply a wide-brush stoke approach. There is either too much or too little metadata management. This simply indicates that not enough effort has gone in to designing and controlling the metadata management. And the cost? Information risk management, either through opportunity loss, information liabilities and / or productivity loss.
Any good data management program must take into account the impact of doing too little, or too much, in controlling their metadata. Ask yourself how much time is spent resolving issues rooted in misunderstandings? How compatible is your metamodel with external standards?
Semantic coherence and metadata cognition are simply modern terms for the classic story of the tower of babel. And I wonder if we will ever learn...
No comments:
Post a Comment